In order to evacuate industrial machines and construction materials in times of flooding at river work sites, it is necessary to perceive the water level increase more than half a day beforehand. In this paper, the authors created a method that predicts the water level up to 1 day ahead by deep learning using meteorological distribution images. The maximum water level of a day at Tsukigata Observatory along the Ishikari River from 2008 to 2020 was predicted. Of the 26 actual flood events, the method was able to predict 17 events 1 day ahead. The method was used to predict the water level of three additional rivers and was able to forecast flood events one day ahead.The technique is be effective for safety management of the river construction.

 This study confirms the mechanism of snowmelt flood estimated from the high-altitude snow depth observation value in the Yoneshiro River basin, and proposes a new method of estimation for snow coverage distribution reflecting the altitude and basin topography in the slope direction. Furthermore, snowmeld flood is evaluated by a hydrological model with particle filters for the one-dimensional unsteady flow model and the snowmelt model improved by snow coverage distribution. Firstly, it is confirmed that temperature and rainfall more than 5 degree centigrade and 5 mm at Moriyoshiyama Ani Ski Resort can predict the high risk flood of the Yoneshiro River. Secondly, the model proposed can ensure the accuracy of the predicted water level 3 hours ahead at the downstream point. It becomes clear that it is important to consider the snowmelt model carrying out flood prediction in the middle and upper basin where the snow coverage is predominant and the influence of snowmelt is significant, and that it is possible to apply the model with only the particle filters except the snow melt model in the downstream regions.

 The main objective of this study is to implement the inexpensive and simple flood-forecast targeting for a mountainous river. We combined the RRI model with a Deep Neural Network (DNN), a type of artificial intelligence technology. A hierarchical neural network is applied to DNN, and the hidden layer is composed by the multiple layers. In this study, the hidden layer was set to three layers, and twenty neurons in each layer. The method of learning for DNN was tested the batch-learning with past floods supervised and the online-machine-learning that updates learning data with current information. The batch-learning was The reproducibility of DNN by the batch-learning was higher than that of the H-Q method. However, a bias error was confirmed in the low water level. In DNN by the online-machine-learning, it was confirmed that high accuracy and stability prediction can be made by learning data higher than the flooding caution water level.

 Our study aimed to build a deep learning model that can predict even flood events caused by an inexperienced rainfall event, for the accuracy improvement of water level predictions at a regulation pond in a drainage pumping station. To improve the deep-learning model accuracy, a physical model created a large amount of pseudo-flood data by virtual rainfalls: 100, 300, and 500 mm for 72 h. The pre-trained model of these flood data was applied to observed data using transfer learning. After extracting only flood events from the observed data, the deep-learning model predicted the largest flood event (TOP 1). The result showed that the pre-trained model with 300 mm/72 h is more accurate than the results of the model without transfer learning. We performed another prediction with the same pre-trained model for the observed data including daily-drainage operations. The result showed that the prediction accuracy was similar to that of non-transfer-learning model as well as the good reproduction of the peak water level for TOP 1.

 In the Hojo River, which flows through central Tottori Prefecture, the flow rate during rainfall increases. Therefore, inundation damage occurs frequently. The river manager judges the river condition by checking the CCTV camera image installed in order to carry out flood control activities promptly when the water overflows. However, there is a shortage of personnel, making continuous monitoring difficult. Therefore, there is a problem that it is not possible to grasp when and where the flood occurred.

 In this study, we developed a model that enables automation of overflow detection using deep learning of artificial intelligence with advanced image analysis ability.In addition, we confirmed the usefulness of the model through field demonstration experiments aimed at improving overflow monitoring. The model developed in this study may be applicable to CCTV cameras installed for river monitoring.

 The Hojo River, which flows through central Tottori Prefecture, frequently causes inundation damage due to increased flow during rainfall. In order to prevent flooding due to rising water levels, river administrators have been operating the diversion weirs installed at the diversion points of the Hojo River drainage canal to reduce the risk of flooding in the backwater sections and the middle and upper reaches of the river. However, continuous day and night monitoring is required for the decision making of weir over-turning, and it is difficult to allocate monitoring personnel and to understand the occurrence of flooding.

 In this study, we developed a water level prediction model to support the decision-making process for the operation of a diversion weir using deep learning, a type of artificial intelligence (AI) technology with advanced time series processing capabilities, and confirmed its usefulness. The method developed in this study can be applied to river management facilities such as weirs, sluices, gates, and flumes, where CCTV cameras are installed to measure the water level.

 In case of heavy rainfall, driftwoods which flow from a mountains area into a river in upper streams of the river frequently block bridges, dams and other river facilities. Driftwood also contributes to secondary damage such as flooding in surrounding areas and damaging bridges. In this study, a sediment failure discrimination model and a driftwood generation prediction model were developed based on numerical data of sediment failure obtained at a heavy rainfall in July 2018. In addition, the sediment failure discrimination model and the driftwood generation prediction model were applied in upstream of the Yasu River in Shiga Prefecture and evaluated damage risks by driftwoods in its tributaries. As a result of evaluating amounts of driftwood flowing into the river between bridges according to several cases of rainfall conditions, some bridges in the Yasu River were shown to be relatively risky due to high potentials of driftwoods and have higher priority for forest and river maintenance.

 Recently, driftwoods disaster has increased by climate change. It was occurred by heavy rainfall event such as the Northern Kyushu Heavy Rain in 2017. In the previous studies, a prediction model for the amount of driftwoods generation has been developed based on the logistic regression analysis. In this study, we analyzed the rainfall tendency based on the Mann-Kendall test, probable maximum rainfall and the amount of driftwood in two periods, that is, before 2017 heavey rain (till 2016) and including the heavy rain (till 2020). The results were obtained as follows: 1) The impact of climate change showed in the rainfall trend after 2017 in the study area; 2) The relationship between the prediction and the potential of driftwoods generation is strongly affected by the trend shift of rainfall due to climate change.

 In this study, the flow of floodwater at the breach point of the Chikuma River during Typhoon No. 19 in 2019 was calculated numerically to obtain the spatial distribution of physical quantities related to the shear stress and driftwood accumulation conditions. The behavior of driftwood was also analyzed based on a particle based model. The accumulation of large wood pieces was observed in areas where the shear stress decreased rapidly. In addition, it was confirmed that the large wood accumulation area corresponds to the point of divergence of the large wood flow path. Based on the buoyancy force and the drag force acting on the large wood piece, the possible locations of vertical accumulation of large wood were calculated, and it was confirmed that the results of the proposed evaluation method were consistent with the actual situation.

 This paper proposed and tested a new large wood retention system in which embayment zone is connected to main channel and woody debris is removed when water level rises. The laboratory flume experiments were conducted by varying the elevation of the embayment zone, entrance shape of embayment zone, installation angle of groyne, groyne position, number of groynes and Froude number. The results showed that the trapping probability is highly influenced by the entrance shape of embayment zone. Then, we measured the flow velocity by PIV and examined the effect of the circulation flow. We also evaluated the followability of driftwood to the water flow varying the dowel model length.

 Under the frequent occurrence of levee-breach by overtopping at flood events, it is necessary to develop an embankment that is hard to be breached. The overtopping water flows along the embankment slope at the beginning, but the flow gradually separates from the slope and becomes to ‘nappe flow’ which further hits and erodes the landside. This study focuses on the erosion phenomenon due to nappe flow by hydraulic experiments and numerical analysis and considers the effect of asphalt pavement on the top of embankment which is made of sandy material. When the nappe flow erodes the landside ground, a large scoured region is formed. The scour collapses the body of embankment, and the pavement on the top of embankment is lost along with the body, causing further erosion. To suppress the broke of the top pavement, it is effective to strengthen the pavement with grid material. This process can be reproduced by numerical model.

 Due to the heavy rain of July in 2018, several levees of the Oda River and its tributaries in the Mabi district of Kurashiki City were breached and serious inundation disaster occurred there. In this area, most causes of the levee breach were due to the overflow, but there were some places where the levees did not break despite the overflow. Since most of the causes of levee breach in Japan are due to overflow, it is very important to predict the bank break due to overflow. In this study, physical quantities that affect the levee breach were selected and examined whether or not the levee breach could be predicted using flood analysis. As a result, it was clarified that the time averaged water level difference between the river side and residence side before the levee broke could be an important index for predicting the levee breach due to overflow.

 Although channel bipolarization has various effects on river channel management and is regarded as a problem, there are very few studies on the relationship between the relative bar height and riverbank erosion based on the data observed in natural rivers. In this paper, we set a bipolarization index (i.e., gravel bar height-bankfull depth ratio; hereinafter, referred to as B-index) and examined the relationship between the bar shape before flooding and riverbank erosion frequency after flooding in four rivers in Hokkaido. Our results show that the width of bank erosion and the frequency of bank erosion tended to increase as the B-index increased. In the channel with alternate bars, the frequency of bank erosion tended to increase sharply at the B-index of 1.0 or higher. Although the sample size of channel with double-row bars is small, the considered sample size shows that the bank is likely to be eroded during floods since the averaged bank height is low relative to the averaged bar height. The results of this study clearly show that the B-index can capture the risk of bank erosion in gravel-bed rivers.

 Sediment Replenishment (SR) is a common restoration measures implemented in Japanese rivers, which aims to recover the degraded river’s morphology and aquatic ecosystem caused by sediment deficit. In this paper, we developed an integral assessment approach to evaluate the effectiveness of SR strategy on flow regime, river bed, and ecological changes along the downstream reaches. Several indices for riverine assessment, such as GUS (Geomorphic Units Survey), HMID (Hydro-morphological Index of Diversity), and Fish Diversity Index (H Value) have been quantified for multiple SR. Key factors as the placed sediment volume, transported rate, and flushing flow discharge of the developed approach has also been successfully applied for Naka River, based on collected data of topography, sedimentation, flow regime, morphology, and water quality. The results reveal that the efficiency of SR was fluctuant for single replenishment at Kohama and additional SR sites should be considered in the future measures of sediment management. Furthermore, the numbers of geomorphic units, such as runs, rapids, and pools were increased around 15% annually, and a tendency of equal distribution of each units can be founded as well. While habitat quality for fish spawning was promoted partly due to the alteration of the morphological and hydrological characteristics (riverbed material, riverbed level, water depth, and flow velocity) at different sites. To summarize, this comprehensive study will show the influences of SR on riverine characteristics and provide valuable recommendations for the operation of sediment replenishment works in the Naka River.

 Fomulas for estimating the maximum local scour depth, which is required for evaluating the stability of the bridge pier. Have been often proposed by experiments under constant water height conditions, and there are few research cases that reproduce the water level history. In this study, we have performed experiments which reproduced the water level history in river in order to verify an effect of repeated water level history on a progress of the local scouring.

 From these results of experiments, we show that local scour depth and local scour length are increased by the water level history, however those are tended to converge by repeating numbers of repetitions. In addition, we show that the local scour length at the upstream of the model is approximately 1.1 to 1.3 times larger than the value estimated from the local scour depth and the angle of repose in water.

 We investigated the long-term changes in river channel by using the observed data from 1947 to 2019 in the middle Tama River. The observed data showed that the mass gravel-digging until 1964 has caused river bed degradation and the river morphological structure with sandbars has been lost. However, since 1999, the bed elevation and width of the main channel began to be controlled mainly by the bed protection works installed at cross-structures and large scale floods over 2,000m³/s. These have caused movements of sandbars sediment supplied to the main channels from the immobilized sandbars due to erosions. The consideration of the long-term changes and responses of river channels provides important technological information for improvements and managements of future river channel.

 In this study, a new experiment on bed variation was conducted to understand the process of stable river channel formation. In addition, effective parameters for estimating revetment failure risk were also investigated. The following were obtained from this study based on the experimental data and actual river data. (1) When a river channel moves toward a dynamic equilibrium state, it approaches the stable channel shape with time. When the river width is constrained, it approaches the stable channel while changing the slope. (2) The stable channel and current channel cross-section comparison are possible to understand characteristics of the river channel variation. (3) The value that indicates longitudinal stable condition and K-values are the effective parameters for estimating revetment failure risk.

 Heavy rainfalls in recent years have increased the number of abutment damage due to river channel migration. Here, we propose a method to statistically estimate bank erosion width using explanatory variables extracted from a physically-based meandering model. We also evaluate the bridge damage risk using the ratio of the half-bridge length to the estimated erosion width. Our model can successfully estimate the erosion width from bed slope, channel curvature, basin area, rainfall intensity, and soil conditions. Also, our study shows that the ratio of the half-bridge length to the estimated erosion width for locations with bridge damage is less than 1.

 Because of a concern of the frequency of intense rainstorm due to global warming, increase of the capacity of retarding basin is required against floods exceeding the present design level is considered, but the space for expansion of existing basins are limited. We examined the feasibility of a retarding basin against excessive flood utilizing an old floodplain in a small scale closed basin using a numerical flow simulation. The study site was a land side plain of a basin of about 300 ha located in the upper reach of Abukuma River, which received an inundation of excessive flood in 2019. Three places were considered for a possible site of an overflow dike, one in the upstream, or two in the mid-stream, and a sluice gate was placed at the downstream end of the basin. The numerical results showed; all three cases reduced about 6-8 percent of the flood discharge exceeding the high water level; and the influence of flooding on houses on the land was minimum by the overflow dike placed at the most downstream. In conclusion, an accumulation of retarding basins of this kind will be feasible as a facility for control of excessive floods, because there are many small closed basins in the upper reach of rivers in Japan.

 Under the global warming condition, it is indispensable to consider flood control measures on the premise of the huge inundation along with bank overtopping or levee breach caused by major floods. In this study, for the Tedori River Basin which is one of leading torrential rivers in Japan, inundation simulations were conducted under the conditions of the assumed maximum rainfall and the rainfall (L1.5 rainfall) between the assumed maximum and the design level to evaluate flood control ability of makeshift flood protection walls installed on existing roads. In both the assumed maximum rainfall and the L1.5 rainfall, the changes in the maximum inundation depth were slight in all three installation patterns of the makeshift flood protection walls. It was clarified that the Tedori River Basin has a steep topographical gradient, so the inundation flow is fast and the facility is difficult to be effective, and the flood control facility has very small water storage capacity.

 In May 2021, the Basic Act on Disaster Control Measures was amended to add "emergency safety measures" that can direct not only horizontal evacuation but also vertical evacuation, but the choice of evacuation action in the stage of "evacuation order" is still left to the residents. Horizontal evacuation may become dangerous in situations where inland flood is occurring. In this study, we conducted flood simulations and inland flood simulations in the middle reaches of the Chikugo River, where inland floods frequently occur, and aimed to propose a mechanism for residents to make safer evacuation decisions by comparing the risks of horizontal evacuation and vertical evacuation. As a result, we found that the risk of horizontal evacuation and vertical evacuation differs greatly in each area, and that it is necessary to instruct appropriate evacuation behavior in each area, such as in some areas where the risk of horizontal evacuation increases faster when inland flood is considered.

 To implement the integrated river plan, it is required to estimate with high accuracy the amount of rainfall-runoff, the discharge capacity of the main river and tributaries, and the risk of inundation. In this study, the flood flow analysis model and the tank model for the upper Tone River system consisting of main river and tributaries were developed to the 2019 flood. The inflows and the discharges of each tributary were estimated from assimilation analysis on the temporal changes in the observed water surface profiles. The water surface profiles and discharge hydrographs in the upper Tone River with discharges from tributaries were explained. The evaluation method for runoff analysis and flood flow analysis based on water level presented in this study is generalizable. Therefore, this method can be applied to other river systems.

 In irrigation ponds, water management protocols, such as reservoir water release, can be achieved during heavy rains using ditch riders. In this study, such water management practices used in O-pond, Kagawa Prefecture, during three frontal rainfall events in 2020 were investigated. Information collected from interviews after the rainfalls revealed that the dich rider operated the inflowing channel diversion gates during heavy rains based on an empirical method to ensure irrigation water storage but also prevent sediment inflow and/or overflow of the irrigation channel. Estimation of the O-pond flood storage rate from reservoir water level records further suggested that the gate-opening condition used impacted the flood storage outcome. Hence, in the O-pond catchment, most rainfall-runoff concentrated in the inflowing channel and, thus, the flood inflow was controlled by the gate operation. In summary, if the runoff pathway is similar for larger rainfall events, the flood control effect of O-pond will depend on the gate operations.

 Paddy Field DAM is designed not to spillover if levee hight is maintained at standard level and runoff via drainage orifice is free flow. However, the results of our numerical simulations at the watershed scale presented a number of paddy fields were spilled over. This was due to several external factors, such as rainwater inflow from adjacent land parcels and backflow into the paddy fields caused by rising water levels in the drainage channels, but it was difficult to identify the factors for each of the many paddy fields in the watershed. In this study, we developed a new analytical method to visualize the inflow from the adjacent land and the backflow from the drainage canal by using the runoff component visualization model to identify the causes of overflowing rice paddies. In addition, it was demonstrated that raising the height of the levees of rice paddies was effective in reducing the occurrence of spillover.

 In this study, we clarified the actual circumstances of the Kuma River flood damage in Hitoyoshi City that occurred on July 4, 2020, and examined the effects of Nakagawara Park and three bridges on abnormal flood currents through hydraulic model experiments. The results obtained showed that the wa-ter level was raised rapidly upstream from the vicinity of Ohashi Bridge under all conditions of flow rates of 6,000 m³/s, 8,000 m³/s and 10,000 m³/s.

 In particular, at the flow rate of 8,000 m³/s, when there is no bridge, the water level will rise by about 2.08 m upstream of the bridge due to the influence of Nakagawara Park alone, and when the bridge girder is submerged, compared to the case without a bridge. It was revealed that the distance is 1.28m di-rectly above the bridge and 2.0m directly above the Mizunote Bridge.

 We estimated the impact of agricultural water utilization facilities (Hyakutaro-mizo and Kono-mizo) in the basin on flooding process. Both waterways run parallel to the Kuma River on the left bank side, and cross the tributaries in various forms such as level crossing, grade separation, and siphon. We developed a model that can express a series of processes of rainfall, runoff, river flow, and flooding.The analysis was performed assuming the case where both channels exist and the case where they do not exist.

 As a result, it was confirmed that the presence of both canals raises the water level in the tributary of the drainage destination from an early stage, and the peak water level of the main river also rises slightly. In addition, while the inundation area decreased around both canals, the inundation was concentrated at the confluence of the main tributaries.When conducting the Integrated Flood Management, the impact of agricultural water use facilities cannot be ignored, and it is desirable to consider it explicitly.

 This study aims to analyze relationship between flood inundation, building damage and human damage in the Kuma River basin due to Reiwa 2 July heavy rain. For this, we conducted field surveys for the height of water mark and situation of building damages. The observed results showed that the outflow of buildings were remarkable in the middle and upper reach of the Kuma River in which the river width was narrow. The results showed that the human damage was caused by the high inundation more than 5 meters and outflow of buildings, suggesting that the vertical evacuation did not work in this disaster scale. Although the inundation started after sunrise, there was a delay in escape, and it was necessary to call early evacuation action before bedtime.

 The projection of variation of rainfall associated with atmospheric fronts in the future climate is critical because torrential rainfall disasters bring structures and livelihood damage all over Japan. In order to investigate the contribution of atmospheric fronts to rainfall, this study found that heavy rainfall disasters occurred when fronts were stagnant in south-western Japan for more than eight days during rainy season from 2006 to 2020. The detection of fronts by physical index is a useful numerical method for identifying the fronts in future climate, and it is important to verify the method. By comparing the detected front using the F diagnostic from atmospheric data and the front depicted on weather maps, it is shown that the physical index used in the F diagnostic increases with the development of an extratropical cyclone with the front.

 Here we discuss the inundation damage and the adaptation of flood control in the downtown area of the Kase River basin using the existing Hokuzan Dam and Kasegawa Dam for the heavy rain in August 2019 and a future heavy rain event on the basis of d2PDF when the global average temperature would increase by 2 °C since the Industrial Revolution. Currently, the Kasegawa Dam is a multipurpose structure with the roles of flood control and water utilization, while the Hokuzan Dam is used only for water utilization. We investigated using the water utilization capacity of the two dams for flood control by using a prior discharge. Our results demonstrated that river water flooding can be controlled in the Kase River basin even under the future heavy rain event in d2PDF as well as the heavy rain in August 2019, and the inundation damage cost decreases about 30 billion JPY if prior discharges from the existing dams are utilized.

 A flood frequency analysis was conducted for the entire river basin of eight rivers in central Japan. The precipitation of d4PDF NHRCM was input to the flood runoff analysis by the storage function method. As a result of analyzing the calculation result, the distribution of small basins, which are closely related to the extreme floods of the main river, differed from basin to basin. The rate of change in flood flow due to climate change was greater for less frequent extreme floods. This method can analyze the flood characteristics of each basin and at the same time analyze the effects of climate change impact.

 In order to evaluate the future change of sediment- and water-related disasters under climate change, this study compared the simulation results of the water and sediment runoff model considering sediment yield by landslides. The output of NHRCM05, a regional climate model, was employed as rainfall inputs. As a result of applying to the Akatani river basin, which was damaged in the heavy rainfall disaster in northern Kyushu, 2017, both the maximum water discharge and total sediment yield volume in the target basin were increased in the future rainfall data. Additionally, the variation of the relationship among total sediment runoff volume and maximum water discharge was also increased. These results implied that the possibility that disasters with a different trend will appear in the future.

 Significant changes and variability in climate are likely to influence the hydrological balance in small and steep Japanese watersheds. Previous studies had confirmed that in the past decades, rapid increase in population and expansion of industrial zones had developed ‘urban heat islands’ in large cities of the country. Hence, this study recognizes the need to analyze the hydroclimatic features of a relatively small, steep, and forested watershed, which lies in the center of Kantō metropolises. Upon applying trend test and determining the rates of change (slope) to the 1965–2015 dataset of the Ogouchi Dam watershed, results showed non-significant trends in rainfall and peak discharge, indicating their high variability in the half-century study period. On the other hand, significantly increasing trends were observed on the annual and seasonal average air temperatures, with the increase in winter being the most drastic. Comparing the rate of temperature increase per month with the rates of increase from its nearby urban stations, local climate in the watershed exhibited milder warming during hot months (Apr–Sep). Meanwhile, the amounts of water stored in the soil for each decade were ably estimated from the rainfall-runoff correlation plots developed during wet season (May–Oct). Results proved a gradual increase in soil water storage capacity, which could be the long-term impact of the continuous efforts of the Government to conserve the forested catchment.

 This study develops a model to simulate a rainfall-runoff-flood inundation model for evaluating the change in flooding pattern in future climate change using a d4PDF dataset. Considering a concept of total flood management in watershed, distributed water balance model (DWBM) in upstream watershed, and two-dimensional river flow and flood inundation model in the Arakawa River (2DRFA) are linked. Reproductivity of the DWBM runoff at the boundary condition (BC) in the main Arakawa River in 2DRFA is good when 150m size grid are used in DWBM, however it is not good in the other BCs in 2DRFA. For future climate, the output by DWBM and Choryu function were applied to the BC in the main Arakawa River and to other BCs, respectively. The present risk also exists in the Yoshimi and Kawajima Town as is similar to the situation after the Arakawa-seisen (river course change), however the risk in main river itself is decreased. In future climate, the inundation water depth in both towns are increased and the inundation pattern around Irumagwa and Arakawa River appears. It is more similar to the situation after the Arakawa-seisen in Edo Era.

 In order to understand the changing trend of the pluvial inundation under the future climate compared to the present, this study focused on comparing inundation simulation results. At first, we picked up rainfall data of Osaka, Sapporo, and Fukuoka from NHRCM2km rainfall data. We calculated the 10-year probability rainfall of the present and future of each city. As a result, larger differences between present and future rainfall appeared in Sapporo and Fukuoka than those in Osaka. These results indicated that the rainfall scale in those cities would become much larger than that of the present. Secondly, we focused on Nakahama district in Osaka city and made four designed rainfalls of present and future from NHRCM2km. We applied them to the simulation model for pluvial inundation. As a result, comparing present and future, the scale of the peak of inundation and the area whose maximum inundated depth was small were estimated to become larger. Thirdly, we analyzed the amount of damage of pluvial inundation. As a result, predicted economic damage in the future became larger than that of the present in every case.

 Future projections, such as d4PDF which is massive ensemble climate prediction data, have uncertainty, but conventional planning methods which produce planned rainfall cannot evaluate the uncertainty. In order to identify future rainfall events that cause damage beyond the planned scale, we conduct pluvial flood simulation using rainfall events generated by both the "conventional planning method" and the "uncertainty-based planning method". Based on the comparison of the maximum waterlogged amount, the future rainfall is categorized and the rainfall patterns that require attention is identified. Furthermore, there is a high correlation between the maximum rainfall for 3-hour and the maximum waterlogging, and it is shown that monitoring the 3-hour maximum rainfall is important for pluvial flood.

 In recent years, rainfall patterns have been changing due to the effects of climate change associated with global warming. In addition, the use of underground spaces such as underground mall and subway is progressing in Osaka, and the subway network is expanding. In this paper, inundation simulations were done for different treatment areas in Osaka City to examine the inundation characteristics of the subway networks. And the inundation risk of subway stations during subway flooding were examined.

 The results show that the subway networks have the highest risk of flooding in each treatment area. In addition, the subway stations that need to be evacuated during pluvial flooding have been clarified.

 A new probabilistic approach is applied to analyze the effect of temporally random variation of rainfall intensity associated with short-term heavy rain on randomness of rainfall inflow into an underground space, in which an approach is further discussed on risks of underground inundation.

 First, a probabilistic model is constructed for random behavior of rainfall intensity of short-term heavy rain by the use of a stationary Gaussian process for describing fluctuation around mean behavior. Then, based upon the probabilistic model, samples of hyetographs are generated on computer, which are transformed into samples of hydrographs through software package InfoWorks ICM. Next, a statistical approach is applied to analyze probability distribution of peak rainfall inflow. The result indicates that Weibull distribution or Gaussian distribution can well reproduce the random distribution of peak rainfall inflow, in comparison with log normal and Gumbel distributions. Further, a systematic approach is proposed on quantifying risk of underground inundation based upon the estimated probability distribution.

 In urban areas of Japan, there are numerous underground spaces, such as subways and shopping malls. These areas hold the possibility of being inundated not only above the ground but also underground when intense rainfall causes pluvial flooding. For this reason, it is necessary to take measures to anticipate urban floods. In this study, an analysis of inundation on the ground surface and underground spaces is performed, and the risk of underground inundation was evaluated using the index of safe evacuation. In addition, as one of the countermeasures against underground inundation, the effect of flood boards on inundation prevention was examined. Furthermore, the factors for improving the success rate of evacuation were discussed by simulating the evacuation of users from an underground mall during inundation. From the results, the inflow to underground spaces is larger as rainfall is greater. Moreover, It is shown that the installation of flood boards reduces the number of inflow points and the amount of inflow, which increases the safe evacuation area of the underground space and improves the success rate of evacuation.

 Recently, inundation damage caused by heavy rain has occurred frequently. Many numerical analysis model on the pluvial inundation were developed. But, observation data used for verification of analysis models are insufficient. In this study, in order to improve the accuracy of numerical analysis, the effects of gutters and attachment sewer pipes were expressed by simple model, and the validity of the analysis model was shown by comparison with observation results. Furthermore, the effect of water level measurement in the sewer system was shown using numerical analysis data. As a result, it is expected that the sewer water level measurement will be used for disaster mitigation measures in places where inundation on the ground appears after the sewer water level rises, and lead time of about 30 minutes was obtained in this study.

 In recent years, the weather has become more extreme on a global scale, and as a result, heavy rainfall has occurred frequently in various places. Even in large cities with advanced sewerage networks, unprecedented heavy rainfall can cause flooding of rivers and flooding of inland waters. In this study, we performed a reproduction calculation using the actual rainfall data of Typhoon No. 19 in 2019 observed by XRAIN of the MLIT in Chofu City, which is the basin of the Tama River.In this study, we performed a reproduction calculation using the actual rainfall data of Typhoon No. 19 in 2019 observed by XRAIN of the MLIT in Chofu City, which is the basin of the Tama River. At the same time, we also examined how the handling of gutter pipes and gutter gates in the target area would affect the rainfall conditions at that time. In addition to this, we also deal with damage that occurred in Ota Ward and Setagaya Ward. This reproduction calculation applies the results of past calculations in the 23 wards of Tokyo, Kawasaki City, and Yokohama City to Chofu City.

 Freshwater fish is one of the major taxonomic groups in river ecosystems and is used as an indicator of river ecosystem health. The knowledge on the distribution and diversity of freshwater fish is essential to develop river management for the conservation of freshwater fishes. In this study, we evaluated the diurnal variations of eDNA concentrations for freshwater fish species in the Saba River, Yamaguchi Prefecture, Japan using environmental DNA (eDNA) quantitative metabarcoding (qMiFish) that enables the quantitative assessment of multiple fish species from water samples. The results showed that the species number, fish community composition, and eDNA concentrations did not significantly differ between daytime and nighttime. These results suggest that researchers can quantitatively understand the fish community compositions from eDNA data by conducting only daytime water sampling in the river. Furthermore, the eDNA concentrations of some fish species have changed between daytime and nighttime, indicating that qMiFish may be able to capture the eDNA dirnal variations with the fish activities.

 In this study, we examined the fish assemblage patterns on different tributary habitat types (natural and artificial tributaries) and relationships between the fish assemblages and environmental factors in the Gonokawa River in Japan using quantitative environmental DNA (eDNA) metabarcoding. The analysis results indicate that the species richness in the natural tributary streams was higher than that in the irrigation channels. In addition, the eDNA concentrations of fishes that tend to prefer sand/gravel substrates or aquatic vegetation in the natural tributary streams were higher than those in the irrigation channels. Further, stream discharge was positively related to certain fishes. This study indicates that quantitative eDNA metabarcoding is an efficient monitoring tool, elucidating that artificial stream modifications would affect fish assemblage compositions in the tributaries.

 The aim of this study is to clarify the limits to the downstream transport distance of environmental DNA (eDNA) containing materials in rivers and the upstream range suitable for grasping biomass of the target species with the eDNA concentrations at a certain point. The study area was the lower reaches of the Takatsu River, western Japan. We simulated one-dimensional dynamics of fish feces and mucus, the primary sources for eDNA containing materials, in several discharge patterns. The downstream transport distance of eDNA varied greatly depending on the discharge; 800 m at the lowest discharge and 1900 m at the highest discharge settings, respectively. The upstream distance suitable for grasping fish biomass in a river varied depending on the longitudinal distribution of the target species and discharge, ranging from 200 to 700 m for Plecoglossus altivelis altivelis and from 50 to 950 m for Candidia temminckii + Opsariichthys platypus. It was also suggested that eDNA can be used to predict the longtitudinal distribution of fish biomass at a higher resolution when the discharge is low.

 Sand cover rate (SCR) is an important factor to explain the foraging habitats of ayu (Plecoglossus altivelis). Hence, a module of riverbed variation in a gravel-bed river with bimodal sediments was developed. The module calculated SCR, and the changes in SCR were defined as being driven by sand transport in the form of a suspended load that is deposited on stones (e.g., cobbles or pebbles) and in the interstices of stones. The prediction accuracy was examined by comparing the calculated SCR with the SCR measured in the field. The results showed that the spatial pattern of sand deposition/erosion predicted by the module was almost close to the measured one, though the predicted SCR tended to be higher/lower than the measured value. Chronological changes in the SCR calculated by the module suggested that sand deposits started forming along shorelines after the peak of a flood and then expanded toward the centers of the flow in channels in the end stage.

 We proposed a method to evaluate the function of a full-section fishway using a 3D point cloud model created from aerial photographs taken by a UAV. In this study, we were able to quickly assess the current condition of the drop-off between fishway pools, which is an important factor in the function of a fishway, and evaluate the potential for fish migration. The area of the pools, the angle of the water vein in each pool and the number of water vein, and the number of inflow and outflow pools, which contribute to securing various migration routes in the full-section fishway, were evaluated for the diversity of upstream routes in the fishway. It was concluded that this method can evaluate the function of the fishway from the viewpoint of the physical environment, and that it is easy to evaluate the ease of fish migration in the fishway.

 In the Kurobe river, cooperative sediment flushing of Dashidaira and Unazuki dams have been executed since 2001 in order to maintain flood control and water utilization system. We studied improvement of dam operational method on the basis of numerical analysis and actual implementation both to reduce the deposited reservoir sediment volume and to lower suspended sediment concentrations just downstream of Unazuki dam. As the numerical analysis, “the preceding drawdown operation of Unazuki dam” seems to be effective where the Dashidaira dam starts to be lowered after the water level of Unazuki dam have been already drawdown to its half of an usual level. As the experiments in 2020, the deposited sediment volume of Unazuki dam in a year was 10,000 m³ equivalent to a 97% reduction and the maximum suspended sediment concentration was 37,000 mg/l equivalent to a 26% reduction in the conventional operation.

 Sediment bypass tunnel (hereafter SBT) is one of the countermeasures for reservoir sedimentation which bypasses inflow sediment to the downstream directly. For operating SBT efficiently, issues to be known are how and when to switch the inflow both for refilling the reservoir water volume and bypassing high concentration sediment, and to find the appropriate sedimentation height in the diversion pool to minimize the sediment inflow to the reservoir. In order that, it is required to predict sediment transport phenomena during SBT operations. In this study, we analyzed the field data at upstream river to predict inflow sediment discharge and created a sediment diversion model at the SBT diversion pool. Based on the comparison of several SBT operation rules, we have proposed the suitable excavation volume of the diversion pool to minimize the amount of sediment flowing into the main reservoir.

 Kamishiiba reservoir has the problem of the turbid water inflow from the upstream catchment area. The main sediment yield is on the bare ground surface after the landslide. In this study, the surface soil and river channel sediments in the catchment area were sampled for examining the similarity with the suspended solid in the turbid water inflow to the reservoir by the mineralogical method. The topsoil of the slope suspected as the cause of the turbidity was aggregated with PIC to examine the possibility of suppressing runoff. In the PIC sprayed area, the soil erosion by raindrop impact can be greatly suppressed, but cracks that developed over time reduced resistance to surface lateral flow, and the sediment yield increases slightly.

 Downstream sediment replenishment has been implemented in some dams. It is important from the viewpoint of continuity of sediment transport, and is also considered to be an effective method to solve the problem of sediment disposal in dams. However, the amount of replenishment is greatly affected by the flood characteristics of each dam, which is an external force.

 In this study, we focused on the discharge data of dams as one of the basic studies on the implementation of downstream sediment replenishment, and compared the flood characteristics of dams that have already taken the sediment disposal measures in this way with those of dams nationwide based on the daily average discharge data. Flood characteristics (number of floods and flood duration) of the dams implementing downstream sediment replenishment were also studied using hourly discharge data.

 Although the calculation of sediment discharge of rivers is very important technique, an accracy of them is not sufficient. In this study, a sediment discharge model of rivers by using filed data and the simple model considering cross-sectional shape of river was proposed by using observation results at the Goryo Brige of Tone River at floods. No-dimensional tractive force was calculated by using the Manning formula. It was slso estimated by using the ADCP data and the Logarythmic law of vertical profile of current velocity. The calculated force was revised by observation results. Furthermore, the revised tractive force was used to calculate the bedload transport ratio. The bedload ratio was also observed by using the bottom track speed of ADCP and the Egashira Flomula. According to other studies, the bedload layer velocity can be estimated by using the bottom track speed of ADCP. The calculated bedload transport ratio by using the revised tractive force was verified by using observation results. Finaly, moored riverboat oscillations were estimated by video analyses and effects of them were removed from bottom track speeds. Comparization between the bedload transport rate by using the conventional model and that by using the revised value considering moored riverboat oscillations revealed the improvement on account of correlation coefficients.